search

taichi-dev / taichi

Productive, portable, and performant GPU programming in Python.
https://taichi-lang.org
Apache License 2.0
25.58k stars 2.29k forks source link

Cannot create ndarray of vec3 and loop over it. #7905

Open mbvaughn21 opened 1 year ago

mbvaughn21 commented 1 year ago

Describe the bug TypeError: Cannot loop over the object <class 'taichi.lang.matrix.VectorNdarray'> in Taichi scope. Only Taichi fields (via template) or dense arrays (via types.ndarray) are supported. How do I make an ndarray of vec3? Why did this type become based on ti.lang.matrix?

To Reproduce Starting from the increment example, I create an array of 3-vectors and attempt to increment the first element of each. Basically, I want perform some computation on every 3-vector in an array of 3-vectors.

import taichi as ti
ti.init()
@ti.data_oriented
class TiArray:
    def __init__(self, n):
        self.x = ti.ndarray(dtype=ti.math.vec3, shape=n)

    @ti.kernel
    def inc(self):
        # Increment the first element of each vector
        for i in self.x:
            self.x[i][0] += 1

a = TiArray(32)
a.inc()

I tried changing to this:

#!/usr/bin/env python3
import taichi as ti

ti.init()

@ti.data_oriented
class TiArray:
    def __init__(self, n):
        self.x = ti.ndarray(dtype=ti.math.vec3, shape=n)

    @ti.kernel
    def inc(self):
        for i in range(self.x.shape[0]):
            self.x[i][0] += 1

a = TiArray(32)
a.inc()

Which gives the following exception:

[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10
[Taichi] Starting on arch=arm64
Traceback (most recent call last):
  File "/Users/ben/test/./test2.py", line 17, in <module>
    a.inc()
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/kernel_impl.py", line 1086, in __call__
    raise type(e)('\n' + str(e)) from None
taichi.lang.exception.TaichiCompilationError: 
File "/Users/ben/test/./test2.py", line 14, in inc:
            self.x[i][0] += 1
            ^^^^^^^^^
Traceback (most recent call last):
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/ast/ast_transformer_utils.py", line 25, in __call__
    return method(ctx, node)
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/ast/ast_transformer.py", line 238, in build_Subscript
    node.ptr = impl.subscript(ctx.ast_builder, node.value.ptr,
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/util.py", line 301, in wrapped
    return func(*args, **kwargs)
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/impl.py", line 164, in subscript
    return value.__getitem__(_indices)
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/util.py", line 309, in wrapped
    assert in_python_scope(), \
AssertionError: __getitem__ cannot be called in Taichi-scope

Which I can fix by passing the array as an arg in this way but I don't understand why it works (and it defeats the purpose of self):

#!/usr/bin/env python3
import taichi as ti

ti.init()

@ti.data_oriented
class TiArray:
    def __init__(self, n):
        self.x = ti.ndarray(dtype=ti.math.vec3, shape=n)

    @ti.kernel
    def inc(self, x: ti.types.ndarray(ti.math.vec3, 1)):
        for i in range(x.shape[0]):
            x[i][0] += 1

a = TiArray(32)
a.inc(a.x)

Log/Screenshots

ben@Bens-M2-Max test % ./test.py
[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10
[Taichi] Starting on arch=arm64
Traceback (most recent call last):
  File "/Users/ben/test/./test.py", line 17, in <module>
    a.inc()
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/kernel_impl.py", line 1086, in __call__
    raise type(e)('\n' + str(e)) from None
taichi.lang.exception.TaichiCompilationError: 
File "/Users/ben/test/./test.py", line 13, in inc:
        for i in self.x:
        ^^^^^^^^^^^^^^^^
Traceback (most recent call last):
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/ast/ast_transformer_utils.py", line 25, in __call__
    return method(ctx, node)
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/ast/ast_transformer.py", line 1352, in build_For
    return ASTTransformer.build_struct_for(ctx,
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/ast/ast_transformer.py", line 1242, in build_struct_for
    impl.begin_frontend_struct_for(ctx.ast_builder, expr_group,
  File "/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/lang/impl.py", line 89, in begin_frontend_struct_for
    raise TypeError(
TypeError: Cannot loop over the object <class 'taichi.lang.matrix.VectorNdarray'> in Taichi scope. Only Taichi fields (via template) or dense arrays (via types.ndarray) are supported.

Additional comments

ben@Bens-M2-Max test % ti diagnose
[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10

*******************************************
**      Taichi Programming Language      **
*******************************************

Docs:   https://docs.taichi-lang.org/
GitHub: https://github.com/taichi-dev/taichi/
Forum:  https://forum.taichi.graphics/

Taichi system diagnose:

python: 3.10.10 (main, Mar 16 2023, 16:52:20) [Clang 14.0.0 (clang-1400.0.29.202)]
system: darwin
executable: /Users/ben/.pyenv/versions/3.10.10/bin/python3.10
platform: macOS-13.2.1-arm64-arm-64bit
architecture: 64bit 
uname: uname_result(system='Darwin', node='Bens-M2-Max.local', release='22.3.0', version='Darwin Kernel Version 22.3.0: Mon Jan 30 20:39:46 PST 2023; root:xnu-8792.81.3~2/RELEASE_ARM64_T6020', machine='arm64')
locale: en_US.UTF-8
PATH: /Users/ben/.pyenv/versions/3.10.10/bin:/opt/homebrew/Cellar/pyenv/2.3.15/libexec:/opt/homebrew/Cellar/pyenv/2.3.15/plugins/python-build/bin:/Users/ben/.pyenv/shims:/opt/homebrew/bin:/opt/homebrew/sbin:/usr/local/bin:/System/Cryptexes/App/usr/bin:/usr/bin:/bin:/usr/sbin:/sbin:/Applications/VMware Fusion.app/Contents/Public:/Users/ben/.cargo/bin:/Users/ben/.pyenv/shims:/opt/homebrew/bin:/opt/homebrew/sbin
PYTHONPATH: ['/Users/ben/.pyenv/versions/3.10.10/bin', '/Users/ben/.pyenv/versions/3.10.10/lib/python310.zip', '/Users/ben/.pyenv/versions/3.10.10/lib/python3.10', '/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/lib-dynload', '/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages']

`lsb_release` not available: [Errno 2] No such file or directory: 'lsb_release'

import: <module 'taichi' from '/Users/ben/.pyenv/versions/3.10.10/lib/python3.10/site-packages/taichi/__init__.py'>

cc: False
cpu: True
metal: True
opengl: False
cuda: False
vulkan: True

`glewinfo` not available: [Errno 2] No such file or directory: 'glewinfo'

`nvidia-smi` not available: [Errno 2] No such file or directory: 'nvidia-smi'
[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10

[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10
[Taichi] Starting on arch=arm64

[W 04/26/23 14:12:26.013 20887529] [misc.py:adaptive_arch_select@772] Arch=[<Arch.opengl: 7>] is not supported, falling back to CPU
[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10
[Taichi] Starting on arch=arm64

[W 04/26/23 14:12:26.502 20887589] [misc.py:adaptive_arch_select@772] Arch=[<Arch.cuda: 5>] is not supported, falling back to CPU
[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10
[Taichi] Starting on arch=arm64

[Taichi] version 1.5.0, llvm 16.0.0git, commit 7b885c28, osx, python 3.10.10

*******************************************
**      Taichi Programming Language      **
*******************************************

Docs:   https://docs.taichi-lang.org/
GitHub: https://github.com/taichi-dev/taichi/
Forum:  https://forum.taichi.graphics/

                                   TAICHI EXAMPLES                                    
 ──────────────────────────────────────────────────────────────────────────────────── 
  0: ad_gravity               25: laplace                 50: physarum                
  1: circle_packing_image     26: laplace_equation        51: poisson_disk_sampling   
  2: comet                    27: mandelbrot_zoom         52: print_offset            
  3: cornell_box              28: marching_squares        53: rasterizer              
  4: diff_sph                 29: mass_spring_3d_ggui     54: regression              
  5: euler                    30: mass_spring_game        55: sdf_renderer            
  6: explicit_activation      31: mass_spring_game_ggui   56: simple_derivative       
  7: export_mesh              32: mciso_advanced          57: simple_texture          
  8: export_ply               33: mgpcg                   58: simple_uv               
  9: export_videos            34: mgpcg_advanced          59: snow_phaseField         
  10: fem128                  35: minimal                 60: stable_fluid            
  11: fem128_ggui             36: minimization            61: stable_fluid_ggui       
  12: fem99                   37: mpm128                  62: stable_fluid_graph      
  13: fractal                 38: mpm128_ggui             63: taichi_bitmasked        
  14: fractal3d_ggui          39: mpm3d                   64: taichi_dynamic          
  15: fullscreen              40: mpm3d_ggui              65: taichi_logo             
  16: game_of_life            41: mpm88                   66: taichi_ngp              
  17: gui_image_io            42: mpm88_graph             67: taichi_sparse           
  18: gui_widgets             43: mpm99                   68: texture_graph           
  19: implicit_fem            44: mpm_lagrangian_forces   69: tutorial                
  20: implicit_mass_spring    45: nbody                   70: two_stream_instability  
  21: initial_value_problem   46: odop_solar              71: vortex_rings            
  22: jacobian                47: oit_renderer            72: waterwave               
  23: karman_vortex_street    48: patterns                                            
  24: keyboard                49: pbf2d                                               
 ──────────────────────────────────────────────────────────────────────────────────── 
42
Running example minimal ...
[Taichi] Starting on arch=arm64
42.0
>>> Running time: 0.20s

Consider attaching this log when maintainers ask about system information.
>>> Running time: 4.84s
bobcao3 commented 1 year ago

Right now NDArray must be passed into the kernel as a kernel argument, it won't be captured through self

feisuzhu commented 1 year ago

Duplicate with #7076